Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Diffusion equation and fick law

  1. May 9, 2006 #1
    Hello every body

    I have previously post my question in this topic:
    Physics Help and Math Help - Physics Forums > Science Education > Homework & Coursework Questions > Other Sciences > Fick and Cottrell Law

    And after Goku suggestion I post my question here.

    So my problem deal with the resolution with fick second law of diffusion (in one dimension)

    In the case of a planar electrode (one dimension) the current density is proportinnal to the concentration of electroactive species: i=-nFkdC(x=0,t)/dt
    From Fick law dC(x,t)/dt=Dd2C(x,t)/d2x

    So in the case of initial condition C(x,t=0)=C0
    I found this solution (not me, on internet) C(x,t)=C0erf(x/(Dt)1/2)

    And so we can deduce Cottrell Law i=-nFAC0(D/Pit)1/2

    Now I would like to found the expression of i in the case of spherical electrode and spherical diffusion, wich species are inside the sphere (yes inside and not outside) of radius R
    With C(R,t)=0 for t>0 and C(r,t=0)=C0
    I would like to found the expression of C(r,t)

    I think fick law in spherical diffusion is dC(r,t)/dt=D1/r2d/dr(r2d/dr(C(r,t)))

    Is it right?

    But now how can I find C(r,t) then dC(r,t)/dr for r=R ???

    Do you have any suggestion?

    Thank you for your attention
  2. jcsd
  3. May 9, 2006 #2


    User Avatar
    Science Advisor
    Gold Member

    I think you're gonna find a lot of standard methods for solving these equation, including Green Functions. Since there is no characteristic scale in your problem, I propose you to solve your problem in terms of a similarity variable [tex]\eta[/tex]. You should be able to work out, via scalement of the equation, that [tex]\eta=r/\sqrt{Dt}[/tex]. That is, you're scaling the radial coordinate with the diffusion length. Performing the change of variable you would end up with an ordinary equation for [tex]\eta[/tex], something like:

    [tex]\frac{-1}{2}\eta\frac{\partial \C}{\partial \eta}=\frac{1}{\eta}\frac{\partial}{\partial \eta}\left(\eta\frac{\partial C}{\partial \eta}\right)[/tex]

    Now it's on your own.
  4. May 9, 2006 #3
    Thank you for your help

    I have done the change you propose
    So now if Iunderstand C(r,t) becomes C(n)

    I found a similar differential equation that the one you propose (n2 instead of n)

    -1/2 n dC/n = 1/n2 d/dn(n2 dC/dn)

    Are you agree with this one?

    And then
    nd2C/dn2 + (2-1/2*n2)dC/dn=0

    Is it a non linear differential equation?

    Is it possible to solve it?
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook